By the second year of life children's learning across a number of domains is mediated by an understanding of others' intentions. The proposed studies will investigate the origins of this foundational system of knowledge. Work conducted in the prior funding period shed light on infants' emerging sensitivity to the intentional structure of action: infants analyze certain actions in terms of their object-directed structure, and this sensitivity seems to be grounded in the details of infants' experience with real-world agents and actions as well as infants' own experience producing goal-directed actions. These conclusions stand in contrast to current approaches assuming that infants possess innate and abstract knowledge about intentions. Rather, infants' initial action knowledge appears to be local and circumscribed. Given this starting point, development would involve the construction of progressively more abstract and general systems of knowledge from these local beginnings. The goal of the proposed studies is to rigorously test this proposal, and thereby gain insight to the processes by which infants become sensitive to the intentional structure of action. The studies will address 2 general questions: (1) Which experiences contribute to infants' sensitivity to the object-directed structure of action? and (2) How broadly do infants generalize information about object-directed action? Series A will explore the effects of self-produced action on infants' sensitivity to the object-directed structure of observed actions. Series B will investigate 2 kinds of observational evidence that have been argued to induce goal attribution in infants. These studies will provide new insights into the emergence of intentional action knowledge. In addition, they may provide insights relevant to the identification and remediation of developmental disorders, ranging from autism to conduct disorder, which involve severe deficits in social reasoning.